1. Field of the Invention
The invention relates to a phase change memory and, in particular, to a programming method for a phase change memory.
2. Description of the Related Art
With rapid development of portable electronic products, demand for nonvolatile semiconductor memories has increased. Phase change memory (PCM) has been considered as one of the leading next-generation nonvolatile semiconductor memory technologies due to some competitive advantages of speed, power saving, capacity, reliability, process integration, and lower cost. Despite of the great potential of PCM technology, the high RESET programming-current requirement makes it difficult to downsize the access transistors and becomes the critical challenge to realize a high density PCM array.
Recently, some methods, such as novel cell structures (edge contact, confined structure, etc.) and new materials (N-doped Ge2Sb2Te5 (GST), O-doped material, etc.), have been developed to resolve the high RESET programming-current issue for PCMs. However, an adverse effect which may accompany with these methods is that the SET programming becomes more and more difficult so that the resistance can not be set as low as desired. As a result, incomplete crystallization state which results in a smaller sensing margin and degraded resistance uniformity may appear in the PCM cell after SET programming. To eliminate the incomplete SET programming, the common strategy is to extend the programming duration of SET current pulse for crystallization, as shown in FIG. 1. By adopting such an approach, a more complete crystallization in the active region can be achieved, thus leading to a larger sensing margin and improved resistance uniformity. However, increasing the pulse width of SET programming pulse also results in increased power consumption, overheating, and degraded speed performance.